Emerging, Experimental & Unproven Tinnitus Treatments

What has actually been tested — and what hasn’t

Last reviewed 24 September 2026 · For anyone who has come across a tinnitus treatment and wants to know where the research stands.

New tinnitus treatments appear all the time — discussed online, sold in shops, offered by clinics, or supported by early laboratory research. Some may turn out to help. Some already have early results in people. Others have been tested carefully and did not do better than a placebo.

This page shows where each claim currently sits in the research process — so you can see what has actually been tested in people with tinnitus, and what hasn’t.

Not proven doesn’t mean proven not to work. It can mean “not tested yet,” “too early to tell,” or “tested without showing benefit” — and those are very different things.

39 treatments and claims, grouped by evidence situation. With JavaScript on, this page adds filters, search and the full evidence detail behind each entry.

No direct human tinnitus evidence found

No studies testing this treatment directly in people with tinnitus were found in our search. Some entries have animal research, or evidence from related drugs or conditions — always shown separately.

• Supplement formulas sold online — Quietum Plus, Cortexi, SonoVive, Silencil, Tinnitus 911, Synapse XT, Sonus Complete
  Evidence situation: No direct human tinnitus evidence
  No study of any of these branded formulas in people with tinnitus was found in our search. Their claims usually rely on studies of individual ingredients or on testimonials.
• Grey-market peptides — BPC-157, ARA-290 (cibinetide), Semax, Selank
  Evidence situation: No direct human tinnitus evidence
  No study of these peptides in people with tinnitus was found in our search, and no animal study of the ear or of tinnitus-like behaviour either. In one Russian report, Semax was one of five treatments given together to people with noise-related hearing loss and tinnitus, so its own effect cannot be separated.
• Stem-cell, exosome & secretome injections — Stem cells, “exosomes,” extracellular vesicles, secretome, cord-blood or bone-marrow cells, “regenerative” clinic injections
  Evidence situation: No direct human tinnitus evidence
  No study of stem-cell, exosome or secretome injections in people with tinnitus was found in our search. The nearest human studies, in hearing loss, measured hearing only. Two registered sudden-hearing-loss trials of vesicles plan a tinnitus questionnaire as a secondary measure.
• Ketamine & esketamine — IV ketamine infusion or injections, ketamine clinics, esketamine, Spravato (esketamine nasal spray), AM-101 (esketamine ear gel — a different formulation)
  Evidence situation: No direct human tinnitus evidence
  No published result shows ketamine infusions affecting tinnitus. A registered ketamine study in people with tinnitus planned tinnitus questionnaires but has posted only brain-chemistry results. Short-lived tinnitus has been reported as a side effect of ketamine in depression trials.
• Psilocybin & other psychedelics — Psilocybin, magic mushrooms, psilocybin retreats, other classic psychedelics (LSD, DMT, 5-MeO-DMT); MDMA shown separately
  Evidence situation: No direct human tinnitus evidence
  No published result shows psilocybin or other classic psychedelics affecting tinnitus. MDMA, a different drug, gave results similar to placebo overall in one very small trial. Some people report tinnitus after psychedelic use.
• Kv7 potassium-channel drugs (in development) — Retigabine (ezogabine) and laboratory analogues, flupirtine, azetukalner (XEN1101), opakalim (BHV-7000)
  Evidence situation: No direct human tinnitus evidence
  In animals, several of these drugs reduced tinnitus-like behaviour when given shortly after loud noise — preventing tinnitus rather than treating it. The only human tinnitus study (flupirtine, 24 people, no comparison group) found no benefit. No tinnitus trial of the newer drugs was found in our search.
• AC102 (in development) — AC102 middle-ear gel injection (AudioCure Pharma)
  Evidence situation: No direct human tinnitus evidence
  In gerbils, AC102 given right after loud noise reduced tinnitus-like behaviour — preventing tinnitus rather than treating it. That study was funded by the company developing AC102. Its only patient trial is in sudden hearing loss, with hearing outcomes; no tinnitus trial of AC102 was found in our search.
• IV NAD+ drips & NAD+ injections — NAD+ infusions or injections
  Evidence situation: No direct human tinnitus evidence
  NAD+ drips have not been tested in people with tinnitus. The closest evidence is a small sudden-hearing-loss trial of low-dose NAD+ injections, where tinnitus was a side-measure with mixed results.
• NAD+ precursor supplements (NR, NMN, nicotinamide) — Nicotinamide riboside (NR), nicotinamide mononucleotide (NMN), nicotinamide (vitamin B3)
  Evidence situation: No direct human tinnitus evidence
  NR and NMN have not been tested in people with tinnitus; a placebo-controlled NMN trial in tinnitus is recruiting in Japan. Nicotinamide, another NAD+ precursor, was tested against placebo in a 1987 trial that found no benefit.

Early human evidence

Small or uncontrolled studies in people with tinnitus. An early hint, not a conclusion.

• PRP ear injections — Platelet-rich plasma (PRP), PRP injections through the eardrum, “regenerative” clinic injections
  Evidence situation: Early human evidence
  PRP injections through the eardrum have four small reports that measured tinnitus. The only comparison found tinnitus scores similar to steroid injections, and none used a placebo.
• Molecular hydrogen (H₂) — Hydrogen inhalation, hydrogen water, H₂
  Evidence situation: Early human evidence
  One small study in China found more improvement when people with recent-onset tinnitus inhaled hydrogen alongside a standard medication. It had no placebo, no independent replication was found in our search, and it tested inhaled hydrogen — not hydrogen water.
• Lipo-Flavonoid — Bioflavonoid (eriodictyol) with B vitamins and vitamin C
  Evidence situation: Early human evidence
  No placebo-controlled test of Lipo-Flavonoid for tinnitus was found in our search. In the only randomized study it was the comparison treatment, and none of the 16 people who finished that arm showed a decrease on the tinnitus questionnaires.
• N-acetylcysteine (NAC) — NAC
  Evidence situation: Early human evidence
  NAC has not been tested as a treatment for established tinnitus in any controlled trial found in our search. In one of two single-patient reports, tinnitus eased while depression also lifted; in the other, NAC alone had not helped. Studies of NAC protecting hearing are a different question.
• Radiofrequency & nerve-block procedures — Pulsed radiofrequency, nerve blocks in the neck (not stellate ganglion block), cervical nerve-root infiltration, occipital-nerve block, cervical plexus block, radiofrequency lesion of the superior cervical sympathetic ganglion, face and ear nerve blocks
  Evidence situation: Early human evidence
  Several different procedures are grouped here, and each has only uncontrolled evidence. Most radiofrequency data come from one clinic, and the first independent series found no change in tinnitus handicap. No study comparing any of these procedures with a sham procedure was found in our search.
• Huneke neural therapy — “Neural therapy” (Huneke), neural therapy after Huneke, procaine or lidocaine injections at skin segments, trigger points, scars and nerve ganglia
  Evidence situation: Early human evidence
  The evidence is one uncontrolled series from 1983, with results described only in words, and one case report with dental treatment at the same time. No study comparing neural therapy with a sham procedure was found in our search. A small randomized trial is registered but not yet recruiting, and its planned dates have passed.
• Upper-cervical chiropractic & neck manipulation — Atlas / C1 adjustment, spinal manipulation, upper-cervical chiropractic
  Evidence situation: Early human evidence
  The evidence is case reports and one retrospective series, in which only 10% of 47 people with neck-related tinnitus improved. No study tested manipulation in tinnitus generally. An association between neck manipulation and rare artery tears has been reported; causality has not been established.
• Vinpocetine & piracetam — Cavinton
  Evidence situation: Early human evidence
  These are two different drugs with separate evidence. No placebo-controlled trial of either for tinnitus was found in our search. Vinpocetine studies were mostly open; two small randomized studies of piracetam, both after sudden hearing loss or noise injury, found no difference from other medicines.
• Japanese Kampo — Kampo herbal formulas; Goshajinkigan (牛車腎気丸), Chotosan (釣藤散), Hangekobokuto (半夏厚朴湯), Saireito (柴苓湯), Yokukansan, Tokishakuyakusan and others
  Evidence situation: Early human evidence
  The only placebo-controlled Kampo trial (hangekobokuto, 12 weeks) found no difference from placebo. Other formulas have only small studies that lacked placebo, adequate randomization or blinding. A result for one formula does not apply to another.

Controlled, not yet confirmed

At least one controlled trial, not yet repeated by an independent research team.

• Botulinum toxin (Botox) — Botox injections into jaw or neck muscles
  Evidence situation: Controlled, not yet confirmed
  In one small placebo-controlled crossover trial (2005), more people rated themselves improved after the toxin than after saline, but other measures did not differ from placebo, and no repeat study was found in our search. A 100-person placebo-controlled trial is underway in Norway.
• Coenzyme Q10 — CoQ10, ubiquinone
  Evidence situation: Controlled, not yet confirmed
  One small placebo-controlled trial (50 people, 6 weeks, taken alongside an antidepressant in both groups) reported lower tinnitus handicap scores. It was registered after recruitment began, some reported outcomes were not registered, and no independent repeat was found in our search.
• Herbal products studied in Iran — Neurotec (tansy, nettle, rose hip), nanocurcumin, a traditional herbal tea, bitter almond oil ear drops
  Evidence situation: Controlled, not yet confirmed
  Four different products, each tested once. Neurotec and a herbal tea did better than placebo on handicap scores at the end of treatment, in trials with important weaknesses. Nanocurcumin did not differ from placebo on its main measure, and almond ear drops added nothing to standard treatment. None has been repeated.
• MemoVigor 2 — MemoVigor 2
  Evidence situation: Controlled, not yet confirmed
  MemoVigor 2 did better than placebo on a tinnitus handicap questionnaire (THI) in one 3-month trial in recent-onset tinnitus, by about 7.7 points on average — a difference between group averages. The trial had manufacturer involvement, and no independent repeat was found in our search.
• AUDISTIM — AUDISTIM Day/Night
  Evidence situation: Controlled, not yet confirmed
  AUDISTIM Day/Night did better than placebo on a tinnitus handicap questionnaire (THI) in one 3-month trial in chronic tinnitus, by about 7 points on average — a difference between group averages, and the plausible range includes small effects. The trial had manufacturer involvement, and no independent repeat was found in our search.
• Stomin A — Stomin A (ストミンA配合錠), nicotinamide + papaverine
  Evidence situation: Controlled, not yet confirmed
  Stomin A’s one trial compared it with plain nicotinamide for 2 weeks; its package insert calls the result significant, while Japan’s guideline and our recalculation do not. It has never been compared with placebo.
• Acamprosate — Acamprosate; glutamate-modulating medicine
  Evidence situation: Controlled, not yet confirmed
  Small placebo-controlled trials of acamprosate were positive, but the largest trial has no published results.

Many studies, but limited placebo or sham evidence

There may be many studies, but adequate placebo or sham comparisons are limited. Without them, expectations and other effects can be difficult to separate from treatment effects.

• Ear acupressure seeds & magnets — Auricular acupressure, ear seeds, Vaccaria seeds, 耳穴压豆, 耳穴贴压, magnetic ear pellets, magnetic ear clips
  Evidence situation: Many studies, limited placebo/sham
  Dozens of Chinese trials report benefit when ear seeds are added to other care, but none compared them with a sham version, and almost none were blinded. A sham-controlled trial is underway in Taiwan, but it tests ear acupressure together with ear nerve stimulation.
• Chinese herbal medicine — 耳聋左慈丸, 通窍活血汤 and other formulas
  Evidence situation: Many studies, limited placebo/sham
  Many Chinese trials report high “effective rates,” but most added herbs to other treatments without a placebo. The one published placebo-controlled trial, in people with hearing loss, reported a positive composite result; no independent repeat was found in our search.
• Stellate ganglion block — SGB, 星状神经节阻滞
  Evidence situation: Many studies, limited placebo/sham
  Eleven randomized trials in a 2026 meta-analysis report better “effective rates,” but none of them used a sham procedure, and the average difference in questionnaire scores between groups was small (about 6 points).

Mixed or conflicting results

Controlled studies disagree with each other.

• Caroverine — Caroverine, Spasmium (Austria); glutamate-modulating medicine
  Evidence situation: Mixed or conflicting
  Caroverine results are mixed: short-term responses after an infusion in some trials, no effect or a fading effect in others.
• TENS & skin electrical stimulation — TENS earpieces, ear-clip TENS, transcutaneous electrical nerve stimulation, cranial electrotherapy (CES, e.g. Alpha-Stim), ear-point microcurrent, RST-Sanexas
  Evidence situation: Mixed or conflicting
  TENS trials conflict: newer small trials report modest questionnaire changes but no loudness change, and older trials of a wearable home device were negative. No tinnitus trial of cranial electrotherapy was found, and most products sold have not been studied themselves.
• Electromagnetic, radiofrequency & magnet devices — Electromagnetic headphones, pulsed electromagnetic field (PEMF), Cleanhearing Sono, ACUFREE, Tinniless, radiofrequency stimulation, ear-canal magnets, Magnesphere
  Evidence situation: Mixed or conflicting
  The devices on this card use different technologies, and their evidence does not transfer. Electromagnetic device trials conflict: one 1-week trial was positive, while placebo trials of a pulsed high-frequency device and of ear-canal magnets were negative. Most products sold have not been studied themselves.

Tested in controlled trials without showing benefit

Available controlled evidence did not show a clear benefit over placebo or the comparison treatment.

• Antinitus patch — “Fractal light” patch
  Evidence situation: Tested without showing benefit
  In the only published placebo-controlled trial (82 people, run by the manufacturer), the main measure — annoyance after 3 weeks — did not differ from placebo. One secondary result at week 7 favoured the patch; handicap scores did not differ. Sweden’s medical products agency issued a marketing ban in 2018, which the company appealed.
• Migraine-medicine combinations (“cochlear migraine”) — Nortriptyline with topiramate, verapamil with paroxetine, migraine-prevention medicines, otologic migraine, cochlear migraine, “cochlear migraine” medication protocols
  Evidence situation: Tested without showing benefit
  In the one placebo-controlled trial, the migraine-medicine groups did not improve significantly more than placebo on the main tinnitus measure. The trial did not select people with migraine. Two further trials are registered; neither has a placebo group.
• Memantine — Memantine; glutamate-modulating medicine
  Evidence situation: Tested without showing benefit
  The better-designed memantine trial found no benefit. A 2025 trial reporting a benefit used alternating allocation and a co-treatment.
• Vitamin B12 — Methylcobalamin, mecobalamin, 甲钴胺
  Evidence situation: Tested without showing benefit
  In the two placebo-controlled trials, B12 did not do better than placebo for tinnitus overall. A possible benefit was seen only within a small group of people with low B12 levels, without a direct comparison against placebo. Correcting a B12 deficiency is a separate medical question.
• Cutting out caffeine — Caffeine withdrawal, decaf, cutting down on coffee or tea
  Evidence situation: Tested without showing benefit
  In a 30-day double-blind crossover trial, Tinnitus Questionnaire scores were essentially the same on caffeinated and decaffeinated days, and people had caffeine-withdrawal symptoms. A separate trial found that a single caffeine dose did not change tinnitus either.
• Homeopathic tinnitus products — Homeopathic drops and tablets sold for “ringing in the ears”
  Evidence situation: Tested without showing benefit
  In the only placebo-controlled trial (28 people, 1998), a homeopathic remedy did no better than placebo. Apart from one small uncontrolled study of Clear Tinnitus supported by its maker, no study of products sold today was found in our search.
• Magnesium (supplements) — Magnesium tablets and capsules
  Evidence situation: Tested without showing benefit
  In the only placebo-controlled trial, people rated their tinnitus about the same after magnesium as after placebo. The trial was small and short, and we found its results only in a trial registry.
• Ozone therapy — Ozone “autohaemotherapy,” oxygen-ozone therapy, ozonated saline, rectal ozone, ozone injections or ear insufflation
  Evidence situation: Tested without showing benefit
  In a controlled trial, ozone blood treatment did no better than no treatment on loudness or handicap scores over 6 months. A second small placebo-controlled trial, known only from a secondary summary, also found no difference. Serious harms, including gas embolism, have been reported with ozone blood treatments.